Entity

Time filter

Source Type

Gabès, Tunisia

Tellili A.,Research Unit MACS | Tellili A.,Higher Institute of Technological Studies at Djerba | Abdelkrim N.,Gabes University | Challouf A.,Gabes University | And 2 more authors.
International Journal of Automation and Control | Year: 2016

This paper deals with the adaptive fault tolerant control for linear time-invariant singularly perturbed systems against actuator failures and external disturbances. By time-scale decomposition using singular perturbation method, the full-order system is decomposed into slow and fast subsystems. An ϵ-dependent fault tolerant controller for the global system is first designed. To avoid numerical stiffness, a simplification based on singular perturbation parameter decoupling is secondly carried out using the reduced subsystems and some manipulations of the Lyapunov equations. The resulting control system is fault tolerant in that, it provides guaranteed asymptotic stability in presence of external disturbances when all control components are operational as well as when actuator failures occur. The stability is guaranteed based on the Lyapunov stability theory provided the singular perturbation parameter is sufficiently small. A numerical example is given to illustrate the proposed method, where the efficiency of the developed approach will be compared with a reliable H∞ control technique. © Copyright 2016 Inderscience Enterprises Ltd.


Jemai W.J.,Research Unit MACS | Jerbi H.,Laboratory LECAP | Abdelkrim M.N.,Research Unit MACS
6th WSEAS International Conference on Dynamical Systems and Control, CONTROL '10 | Year: 2010

The present work proposes a new method to the problem of implementation of Input-State feedback linearization. Such method is based on the search of a diffeomorphism which transforms the original nonlinear system into a linear one in the controllable canonical form with an external reference input, and the subsequent using of linear pole-placement techniques. The problem is solved without need of the differential geometric complexity of the feedback linearization technique. It is based however, on an analytical method using the development in to generalized Taylor series of vectorial nonlinear functions and Kronecker product tools. The algebraic developments, we present in this paper, use the Monte Carlo optimization method to choose the best parameters of the polynomial feedback control in order to ensure the effectiveness of the diffeomorphism. A simulation study is synthesized to show the effectiveness of the proposed approach.


Jemai W.J.,Research Unit MACS | Jerbi H.,Laboratory LECAP | Abdelkrim M.N.,Research Unit MACS
WSEAS Transactions on Systems and Control | Year: 2010

In the present work, we propose a novel polynomial approach to approximate the Input-State feedback linearization control. The aim of this new method is to simplify the implementation complexity of the exact Input-State feedback linearization. Indeed, the present approach leads to an analytical control law via analytical nonlinear transformations without need to resolve a set of partial differential equations. In fact, the analytical control law, determined via the proposed work, is dependent to an arbitrary choice of some parameters. So and in order to ensure a satisfactory evolution of the control input, we resort to optimization methods to have the optimal values of parameters. A study simulation is presented to show the effectiveness of the proposed approach.


Tellili A.,Research Unit MACS | Tellili A.,Higher Institute of Technological Studies at Djerba ISETJB | Abdelkrim N.,Research Unit MACS | Abdelkrim N.,Gabes University | And 4 more authors.
Information Technology and Control | Year: 2015

This work presents an adaptive approach for fault tolerant control of singularly perturbed systems, where both actuator and sensor faults are examined in presence of external disturbances. For sensor faults, an adaptive controller is designed based on an output-feedback control scheme. The feedback controller gain is determined in order to stabilize the closed-loop system in the fault free case and vanishing disturbance, while the additive gain is updated using an adaptive law to compensate for the sensor faults and the external disturbances. To correct the actuator faults, a state-feedback control method based on adaptive mechanism is considered. The both proposed controllers depend on the singular perturbation parameter ε leading to ill-conditioned problems. A well-posed problem is obtained by simplifying the Lyapunov equations and subsequently the controllers using the singular perturbation method and the reduced subsystems yielding to an ε-independent controller. The control scheme, designed based on the Lyapunov stability theory, guarantees asymptotic stability in presence of additive faults and external disturbances provided the singular perturbation parameter is sufficiently small. Finally, a numerical example is presented to demonstrate the effectiveness of the obtained results. © 2015, Kauno Technologijos Universitetas. All rights reserved.


Trab S.,Research Unit MACS | Bajic E.,University of Lorraine | Zouinkhi A.,Research Unit MACS | Abdelkrim M.N.,Research Unit MACS | And 2 more authors.
Procedia Computer Science | Year: 2015

One of the most relevant topics in warehouse management system (WMS) is the security issue and concerns the optimal placement allocation of products with respect of product and human safety in a sustainable system. Knowing that differences often exist between virtual view of products placements in centralized WMS and the real situation in the facility due to unplanned movements resulting from human errors or products' misplacements, we propose a reactive and compatibility constraint approach for product storage allocation. Our aim is to reduce the size of floating locations largely used in WMS and to avoid the inherent risks of hazardous accidents which can be generated by incompatibility between products and then to minimize the total logistic costs and to guarantee higher warehousing service levels in a safety monitored environment. This work proposes a multi-agent architecture for product allocation planning with compatibility constraints (PAP/CC), which uses a decision mechanism for product's placement, based on negotiations between agents associated to compatibility tests. This approach represents an improvement key for decentralized management of warehouses in a dynamic and reactive environment. Negotiations mechanisms relying on an Internet of Things (IoT) infrastructure and multi agent systems are defined in order to solve security problem of product allocation operations. Industrial deployment of IoT platform represents an ideal solution for decentralized management and to support collaboration between products and shelves. A simulation case of the proposed interactions mechanisms is provided with the use of NetLogo environment which offers many advantages to control agents and to describe their interactions in a graphical environment. © 2015 The Authors.

Discover hidden collaborations